8 8
Humberg

Pine H64 Model B

Recommended Posts

i do not know if there are any Model A users around ;) i have just build an  image Armbian_5.77_Orangepi3_Ubuntu_bionic_dev_5.0.1_desktop and booted up in Desktop environment.  I would not recommend this board for such  use-case but it is nice to see that development is coming up. I am going to test network and USB3 later this day. 

Share this post


Link to post
Share on other sites

Built Armbian_5.77_Pineh64_Ubuntu_bionic_dev_5.0.2.img today.

 

Looked to have a clean boot, working video, full 3gb ram available, Ethernet working, Wi-fi working.

 

Can play with it more later, but it is looking pretty good.

Share this post


Link to post
Share on other sites
On 3/16/2019 at 1:54 PM, Da Alchemist said:

i do not know if there are any Model A users around ;) i have just build an  image Armbian_5.77_Orangepi3_Ubuntu_bionic_dev_5.0.1_desktop and booted up in Desktop environment.  I would not recommend this board for such  use-case but it is nice to see that development is coming up. I am going to test network and USB3 later this day. 

 

I have a model A,  and i am interesting to test the image.  Juste tell me where a i can dowload it. Thanks in advance

Share this post


Link to post
Share on other sites

I have build this Image with Armbians Build system. I am living on the german coutryside, my internet connection is far to slow to upload the image. Perhaps @Igor can change some things, so that there will be nightly builds with  5.01 Kernel for pineH64 again

Share this post


Link to post
Share on other sites
1 minute ago, Da Alchemist said:

I have build this Image with Armbians Build system. I am living on the german coutryside, my internet connection is far to slow to upload the image. Perhaps @Igor can change some things, so that there will be nightly builds with  5.01 Kernel for pineH64 again

it would be great if it's possible

Share this post


Link to post
Share on other sites

Hi,

Compiled yesterday evening a dev (kernel 5.0.5) image for pineh64 and run it on my B model (writting this post with it). It works quite well (hdmi, wifi, eth, desktop).

Audio does not work (only a dummy output) and usb3 port also has some issues.

In case you want more details, here is the output of armbianmonitor -u: http://ix.io/1Fd8

 

Share this post


Link to post
Share on other sites

I've just build Bionic dev 5.0.5 for the PineH64. Everything seems to work immediatly. Nice surprise. Only no on-board sound.
I am wondering why it can't clock to 1.8Ghz. It is actually more useable at 1.5Ghz since it doesn't overheat without heatsink+fan. But it would be nice to have the possibility.

Share this post


Link to post
Share on other sites
On 3/29/2019 at 3:10 PM, Da Alchemist said:

can change some things, so that there will be nightly builds with  5.01 Kernel for pineH64 again

I've turn ON nightly BETA of PineH64 for next evening, but I won't leave it for daily basis ...

Share this post


Link to post
Share on other sites

@martinayotte Reboot sometimes fails. It crashes with only the cursor visible(not blinking...)  Looks like the moment it should receive the reboot command. Maybe related with the OPi problem? A timing issue or so?
 

Share this post


Link to post
Share on other sites
59 minutes ago, martinayotte said:

Did you had a Serial-TTL attached while it failed ? 

I'll do that tomorrow.

Share this post


Link to post
Share on other sites
(edited)

tried mine, boots, runs smoothly, capped at 1488 MHz as described elsewhere, I got an IRQ 95 disabled and a bunch of regulator not founds

 

@martinayotte I found it, the dmesg gripes of the voltage for the opp setting not being supported, so it disables 1.8 GHz.

 

reg_dcdca does show 1080000 uV as it's max in the dt.  Can that go higher, or can we hazard changing the 1.8 GHz opp from

 

		opp-1800000000 {
                       opp-hz = /bits/ 64 <1800000000>;
                       opp-microvolt = <1160000>;
                       clock-latency-ns = <244144>; /* 8 32k periods */
               };

to

		opp@1800000000 {
                       opp-hz = /bits/ 64 <1800000000>;
                       opp-microvolt = <1080000 1080000 1160000>;
                       clock-latency-ns = <244144>; /* 8 32k periods */
               };

Doing some other general cleaning, I don't think it will make anything new work (maybe bluetooth?), but you'll see fewer rejects on compile.

Edited by TonyMac32

Share this post


Link to post
Share on other sites
41 minutes ago, TonyMac32 said:

capped at 1488 MHz as described elsewhere

Could you point me to that? It seems reasonable. Same performance as clocked at 1.8Ghz without a fan since the throttling. These Allwinner boards are not made for high-performance, but useabillity.
To be able to use it without a heatsink+fan without constant throttle is a good trade. I also think the SoC is too small at 15mm for 1.8Ghz. Too little room to dissipate the heat well. On PineH64 that's a lot better at 1.5Ghz.
I've odered 15mm copper shims and bigger heatsink. I should be able to fix it onto the opi3 in 2 wholes. (also for a video, cooling sbc's where I replace the thermal pad of the M4 with a shim)

For the PineH64 I can see a few spots they'd could be usefull. I hope the Pine is more power concervative. 0.5A iddle is too much when you're running from power banks. The OPi3 its great advantage is 5Ghz wifi. For long I didn't care. But now I don't want 2.4Ghz anymore.
But with the N2 on the way, a lot of other boards will have to shift to other tasks,

Share this post


Link to post
Share on other sites
2 minutes ago, NicoD said:

I also think the SoC is too small at 15mm for 1.8Ghz.

Don't you have an XU4?  :-P

 

I saw written somewhere someone was mentioning it would only go to 1.488, I was just confirming, and found a good reason why.  my dmesg looks a lot cleaner, I eliminated some patches that were no longer needed but written a bit differently than the actual mainline code, so some of them half-applied.  :blink:

Share this post


Link to post
Share on other sites
17 minutes ago, TonyMac32 said:

Don't you have an XU4:-P 

Yes. Doesnt it have the same problem at high clocks? I've got the big heatsink, and a big fan, and it can't be kept cool enough not to throttle. 1.8Ghz is ok. This over a long +1 hour maxed out task. I learned there that a lower clock can be faster when it doesn't throttle.
It's of course not as bad as the XU4. That heatsink would keep it cooled passively.

Share this post


Link to post
Share on other sites
1 minute ago, NicoD said:

Yes. Doesnt it have the same problem at high clocks? I've got the big heatsink, and a big fan, and it can't be kept cool enough not to throttle.

I got a GPU RAM water block on mine...  :lol::ph34r:  It gets to the mid 60's even then.

Share this post


Link to post
Share on other sites

OK, OPi3 shows dcdca going to 1160000.  The datasheet on the AXP805 has a weird 2-stage step size, but more concerning to me is the fact it seems to suggest there are 72 steps in a  6-bit number... :wacko:

 

I'll make it match Opi-3 and not mess with the opp other than to correct it's syntax to match the rest of the opps

 

@NicoD even sitting at 1.488 forever running youtube video it isn't even warming up with a Tinker factory heatsink on it.  I think we're probably in good shape heat wise, surprisingly.

 

Oh my goodness this is a nice little board...

Share this post


Link to post
Share on other sites
10 hours ago, NicoD said:

Reboot sometimes fails. It crashes with only the cursor visible(not blinking...) 

I observe quite the same behavior, but if I remember well, the cursor was blinking and after about 1 minute the board reboots.

Share this post


Link to post
Share on other sites
4 hours ago, jeanrhum said:

after about 1 minute the board reboots.

Did you have a Serial-TTL attached to see what is happening during that minute ?

I bet it is the NetworkManager that is waiting to be shutdown ...

Share this post


Link to post
Share on other sites
7 hours ago, TonyMac32 said:

even sitting at 1.488 forever running youtube video it isn't even warming up with a Tinker factory heatsink on it.  I think we're probably in good shape heat wise, surprisingly. 

Ideed, even without a heatsink it beahaves nice at 1.5Ghz. Just not into throttle temps.
That's why I think for the OPi 1.5Ghz might also be a more senseable frequency. I'll try it out today without a heatsink. At 1.8Ghz it's a worse experience if you have slowdowns of the system constantly.
I like it too. It's simple and compact. Seems to do what it must do. I had no problems yet, installed onto eMMC. Runs great. If there were a bit more steps, then maybe 1.6Ghz would be ok, or a little lower voltage maybe . The voltage must be a lot higher to reach 1.8Ghz. The difference in how quicky it rises is just immence with 1.5Ghz. In less than a second it can rise from 50C to 80C then. That's just a hotspot that can't release it's heat. So I think either 1.8Ghz is truly it's limit and they made it work on the OPi somehow badly.
I'll set up a small display and sbc to read the ttl.

 

Share this post


Link to post
Share on other sites
39 minutes ago, martinayotte said:

Did you have a Serial-TTL attached to see what is happening during that minute ?

I will look at this when coming back home.

Share this post


Link to post
Share on other sites
21 minutes ago, NicoD said:

At 1.8Ghz it's a worse experience if you have slowdowns of the system constantly.

I ran it at 1.8 for about 30 minutes playing videos and keeping it clocked up, I never exceeded 65 C.  That is with the tiny Tinker heat sink, so I think it is ok.  

Share this post


Link to post
Share on other sites
6 hours ago, TonyMac32 said:

I ran it at 1.8 for about 30 minutes playing videos and keeping it clocked up, I never exceeded 65 C.  That is with the tiny Tinker heat sink, so I think it is ok.  

Is that with the Orange Pi3? I guess it is since I can't clock the H64 to 1.8Ghz.
I've got higher temperatures with the OPi3 with a fan at 3.3V+heatsink. I do have to say that the double sided tape used is very bad. Waiting on new heatsinks to arrive.
I always use Blender for this test since it's using near 100% for a long time. It does take some time before it heats up.

I also have seen different temperature readings on the OPi3. Still no idea what that was. Sometimes it showed in 50°C, sometimes in 60°C, sometimes in 70°C. With the same tasks, just a reboot difference.
It did throttle then at 65°C instead of 75°C. So everything shifted.
I also tried without a heatsink on the OPi3 and it was very problematic. More throttle time than 1.8Ghz while doing anything longer than a few seconds maxed out.
Here my temps with OPi3 with heatsink and no fan. It was a lot worse without heatsink.

Spoiler

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:27:10: 1800MHz  0.23  11%   1%   7%   0%   2%   0%  49.4°C
16:29:10: 1800MHz  1.04  33%   1%  32%   0%   0%   0%  61.9°C
16:31:10: 1800MHz  1.22  34%   1%  32%   0%   0%   0%  61.9°C
16:33:10: 1800MHz  1.22  29%   0%  28%   0%   0%   0%  61.5°C
16:35:10: 1800MHz  1.13  29%   0%  29%   0%   0%   0%  61.3°C
16:37:10: 1800MHz  1.07  29%   0%  28%   0%   0%   0%  61.4°C
16:39:10: 1800MHz  1.25  29%   0%  28%   0%   0%   0%  61.9°C

System health while running OpenSSL benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:40:09: 1800MHz  1.23  20%   1%  18%   0%   1%   0%  62.2°C
16:40:19: 1800MHz  1.19  29%   0%  28%   0%   0%   0%  64.8°C
16:40:29: 1800MHz  1.16  29%   0%  28%   0%   0%   0%  67.0°C
16:40:39: 1800MHz  1.14  29%   0%  28%   0%   0%   0%  68.7°C
16:40:49: 1800MHz  1.20  28%   0%  28%   0%   0%   0%  68.8°C
16:40:59: 1800MHz  1.17  29%   0%  28%   0%   0%   0%  70.2°C
16:41:09: 1800MHz  1.21  28%   0%  28%   0%   0%   0%  68.9°C
16:41:19: 1800MHz  1.18  28%   0%  28%   0%   0%   0%  70.2°C
16:41:29: 1800MHz  1.15  28%   0%  27%   0%   0%   0%  69.4°C
16:41:39: 1800MHz  1.13  28%   0%  27%   0%   0%   0%  69.0°C
16:41:49: 1800MHz  1.11  28%   0%  28%   0%   0%   0%  70.2°C

System health while running 7-zip single core benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:41:57: 1800MHz  1.10  21%   1%  18%   0%   1%   0%  70.8°C
16:42:57: 1800MHz  2.64  28%   0%  27%   0%   0%   0%  63.8°C
16:43:57: 1800MHz  2.88  29%   1%  28%   0%   0%   0%  63.8°C
16:44:57: 1800MHz  3.13  29%   1%  28%   0%   0%   0%  63.1°C

System health while running 7-zip multi core benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:45:44: 1800MHz  3.28  22%   1%  19%   0%   1%   0%  64.6°C
16:46:06: 1320MHz  3.41  84%   1%  82%   0%   0%   0%  77.5°C
16:46:27: 1320MHz  3.87  92%   1%  90%   0%   0%   0%  77.5°C
16:46:48: 1800MHz  3.81  89%   2%  87%   0%   0%   0%  74.6°C
16:47:08: 1320MHz  4.15  87%   1%  85%   0%   0%   0%  75.3°C
16:47:28: 1320MHz  4.10  92%   1%  91%   0%   0%   0%  72.8°C
16:47:48: 1800MHz  4.06  90%   2%  88%   0%   0%   0%  68.8°C
16:48:09: 1320MHz  3.98  91%   2%  88%   0%   0%   0%  70.2°C
16:48:30: 1800MHz  4.04  86%   1%  85%   0%   0%   0%  70.3°C
16:48:50: 1800MHz  4.17  90%   1%  89%   0%   0%   0%  70.4°C
16:49:10: 1800MHz  4.22  89%   2%  87%   0%   0%   0%  76.1°C
16:49:31: 1800MHz  4.45  95%   2%  92%   0%   0%   0%  71.1°C

System health while running cpuminer:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:49:45: 1800MHz  4.48  29%   1%  27%   0%   0%   0%  70.9°C
16:50:07: 1320MHz  4.57  99%   0%  98%   0%   0%   0%  82.3°C
16:50:29: 1800MHz  4.81 100%   0%  99%   0%   0%   0%  76.1°C
16:50:51: 1800MHz  4.59 100%   0%  99%   0%   0%   0%  73.8°C
16:51:12: 1800MHz  4.42 100%   0%  99%   0%   0%   0%  74.0°C
16:51:34: 1320MHz  4.45 100%   0%  99%   0%   0%   0%  83.3°C
16:51:56: 1800MHz  4.35 100%   0%  99%   0%   0%   0%  77.2°C
16:52:17: 1320MHz  4.25 100%   0%  99%   0%   0%   0%  83.1°C
16:52:39: 1320MHz  4.25 100%   0%  99%   0%   0%   0%  74.5°C
16:53:01: 1320MHz  4.46 100%   0%  99%   0%   0%   0%  74.8°C
16:53:23: 1800MHz  4.33 100%   0%  99%   0%   0%   0%  83.7°C
16:53:44: 1320MHz  4.30 100%   0%  99%   0%   0%   0%  77.7°C
16:54:06: 1320MHz  4.34 100%   0%  99%   0%   0%   0%  75.3°C
16:54:28: 1320MHz  4.40 100%   0%  99%   0%   0%   0%  77.6°C

 

Here the same with 3.3V fan on. Good enough to not throttle.

Spoiler

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
19:27:00:     MHz  1.54   9%   1%   7%   0%   1%   0%  42.1°C
19:29:00:     MHz  1.28  34%   1%  32%   0%   0%   0%  52.0°C
19:31:00:     MHz  1.51  34%   1%  32%   0%   0%   0%  46.4°C
19:33:00:     MHz  1.25  29%   0%  29%   0%   0%   0%  45.2°C
19:35:00:     MHz  1.12  30%   0%  29%   0%   0%   0%  45.8°C
19:37:00:     MHz  1.30  29%   0%  28%   0%   0%   0%  45.3°C
19:39:00:     MHz  1.34  29%   0%  28%   0%   0%   0%  45.1°C

System health while running OpenSSL benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
19:39:51:     MHz  1.18  14%   1%  12%   0%   0%   0%  45.2°C
19:40:01:     MHz  1.15  29%   0%  28%   0%   0%   0%  50.2°C
19:40:11:     MHz  1.13  29%   0%  28%   0%   0%   0%  47.3°C
19:40:21:     MHz  1.11  28%   0%  28%   0%   0%   0%  47.4°C
19:40:31:     MHz  1.09  29%   0%  28%   0%   0%   0%  49.7°C
19:40:41:     MHz  1.16  28%   0%  27%   0%   0%   0%  51.3°C
19:40:52:     MHz  1.29  28%   0%  28%   0%   0%   0%  47.4°C
19:41:02:     MHz  1.25  28%   0%  28%   0%   0%   0%  47.9°C
19:41:12:     MHz  1.21  29%   0%  28%   0%   0%   0%  48.0°C
19:41:22:     MHz  1.18  29%   0%  28%   0%   0%   0%  50.1°C
19:41:32:     MHz  1.15  29%   0%  28%   0%   0%   0%  51.4°C

System health while running 7-zip single core benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
19:41:39:     MHz  1.14  15%   1%  13%   0%   0%   0%  51.7°C
19:42:39:     MHz  2.53  29%   1%  28%   0%   0%   0%  46.0°C
19:43:40:     MHz  3.17  30%   1%  29%   0%   0%   0%  46.0°C
19:44:40:     MHz  2.79  30%   1%  28%   0%   0%   0%  45.8°C

System health while running 7-zip multi core benchmark:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
19:45:28:     MHz  3.01  16%   1%  14%   0%   0%   0%  46.8°C
19:45:49:     MHz  3.40  85%   1%  84%   0%   0%   0%  63.8°C
19:46:09:     MHz  3.46  89%   2%  87%   0%   0%   0%  52.5°C
19:46:31:     MHz  3.88  94%   2%  92%   0%   0%   0%  58.9°C
19:46:52:     MHz  4.26  87%   1%  85%   0%   0%   0%  64.4°C
19:47:12:     MHz  3.75  89%   2%  86%   0%   0%   0%  60.0°C
19:47:33:     MHz  3.52  90%   2%  87%   0%   0%   0%  59.6°C
19:47:53:     MHz  4.25  99%   2%  96%   0%   0%   0%  59.5°C
19:48:13:     MHz  4.34  82%   1%  80%   0%   0%   0%  60.4°C
19:48:34:     MHz  4.18  93%   2%  90%   0%   0%   0%  59.5°C
19:48:54:     MHz  4.13  90%   2%  88%   0%   0%   0%  60.5°C

System health while running cpuminer:

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
19:49:14:     MHz  4.47  21%   1%  19%   0%   0%   0%  53.6°C
19:49:35:     MHz  4.34  99%   1%  98%   0%   0%   0%  71.6°C
19:49:57:     MHz  4.30 100%   0%  99%   0%   0%   0%  72.3°C
19:50:18:     MHz  4.42 100%   0%  99%   0%   0%   0%  72.9°C
19:50:40:     MHz  4.30 100%   0%  99%   0%   0%   0%  73.4°C
19:51:02:     MHz  4.28 100%   0%  99%   0%   0%   0%  73.1°C
19:51:23:     MHz  4.43 100%   0%  99%   0%   0%   0%  73.7°C
19:51:45:     MHz  4.31 100%   0%  99%   0%   0%   0%  73.8°C
19:52:07:     MHz  4.22 100%   0%  99%   0%   0%   0%  74.0°C
19:52:29:     MHz  4.42 100%   0%  99%   0%   0%   0%  73.5°C
19:52:50:     MHz  4.30 100%   0%  99%   0%   0%   0%  73.8°C
19:53:12:     MHz  4.28 100%   0%  99%   0%   0%   0%  73.4°C
19:53:34:     MHz  4.42 100%   0%  99%   0%   0%   0%  73.6°C
19:53:55:     MHz  4.30 100%   0%  99%   0%   0%   0%  73.8°C

 

I've replaced my power cable for these tests since that wasn't a good one. It was from OPi, USB to barrel jack. Up to 1V voltage loss on USB ports with that. When I use a microUSB I've got a stable voltage always above 5V. Could be the temperature differences was because of this. Didn't see it the last days with it.

I'm not seeing any of these things on the H64.

Now testing temps of the H64 without heatsink. First minutes it stayed around 65°C. After 10 minutes it's 70°C. Now (16min) at 73°C. Maybe it will reach it's throttle temp, but it will not be throttling much at 1.5Ghz.(it did reach 75°C after 19m20s) It clearly throttled a bit, temperature is suddenly 3°C less and again very slowly rising to 75°C. Maybe 0.05% throttle time. That's about perfect I'd say.

I did try the OPi3 image on the H64 yesterday. It seemed to boot fine, but USB wasn't working. I think I remember it was 1.8Ghz.

Share this post


Link to post
Share on other sites

Time        CPU    load %cpu %sys %usr %nice %io %irq   Temp
16:49:45: 1800MHz  4.48  29%   1%  27%   0%   0%   0%  70.9°C
16:50:07: 1320MHz  4.57  99%   0%  98%   0%   0%   0%  82.3°C
16:50:29: 1800MHz  4.81 100%   0%  99%   0%   0%   0%  76.1°C
16:50:51: 1800MHz  4.59 100%   0%  99%   0%   0%   0%  73.8°C
16:51:12: 1800MHz  4.42 100%   0%  99%   0%   0%   0%  74.0°C
16:51:34: 1320MHz  4.45 100%   0%  99%   0%   0%   0%  83.3°C
16:51:56: 1800MHz  4.35 100%   0%  99%   0%   0%   0%  77.2°C
16:52:17: 1320MHz  4.25 100%   0%  99%   0%   0%   0%  83.1°C
16:52:39: 1320MHz  4.25 100%   0%  99%   0%   0%   0%  74.5°C
16:53:01: 1320MHz  4.46 100%   0%  99%   0%   0%   0%  74.8°C
16:53:23: 1800MHz  4.33 100%   0%  99%   0%   0%   0%  83.7°C
16:53:44: 1320MHz  4.30 100%   0%  99%   0%   0%   0%  77.7°C
16:54:06: 1320MHz  4.34 100%   0%  99%   0%   0%   0%  75.3°C
16:54:28: 1320MHz  4.40 100%   0%  99%   0%   0%   0%  77.6°C

Here it constantly jumped from 85°C to 73°C. Throttle went on, and directly it's 73°C, throttle goes off and immediatly it's 85°C. So this isn't good. Could be my board... I've got some that just have their own will.

Share this post


Link to post
Share on other sites
1 minute ago, NicoD said:

guess it is since I can't clock the H64 to 1.8Ghz.

Nope, I have the H64 running 1.8 now.  It was the dcdca max voltage setting in the H64 dts, it was set to 1.06 volts.  I made it match the OPi3 at 1.16.  the driver was kicking out the 1.8 GHz opp before because the opp requires 1.16.  

Share this post


Link to post
Share on other sites
1 minute ago, TonyMac32 said:

Nope, I have the H64 running 1.8 now.  It was the dcdca max voltage setting in the H64 dts, it was set to 1.06 volts.  I made it match the OPi3 at 1.16.  the driver was kicking out the 1.8 GHz opp before because the opp requires 1.16.   

Oh, great.
 

 

Just now, TonyMac32 said:

If you build an H64 dev image now you can try it out

I guess I'll start doing that. I'm wondering if it's the same behaviour as the OPi3. Thanks you....

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
8 8